How to extract structured drilling report data from PDF into JSON using Python?

[PentapatiAdarsh]

I’m building a RAG-style application and I want to extract data from PDF reports into a structured JSON format so I can send it directly to an LLM later, without using embeddings.

Right now I’m:

describing the PDF layout in a YAML pattern,

using pdfplumber to extract fields/tables according to that pattern,

saving the result as JSON.

On complex reports (example screenshot/page attached), I’m running into issues keeping the extraction 100% accurate and stable: mis-detected table rows, shifted columns, and occasional missing fields.

My questions:

Are there better approaches or libraries for highly reliable, template-based PDF → JSON extraction?

Is there a recommended way to combine pdfplumber with layout analysis (or another tool) to make this more robust and automatable for RAG ingestion?

Constraints:

Reports follow a fixed layout (like the attached Daily Drilling Report).

I’d like something that can run automatically in a pipeline (no manual labeling).

Any patterns, tools, or example code for turning a fixed-format PDF like this into consistent JSON would be greatly appreciated.
report2.pdf